AI Threatens to Turn Advertising Into a Toll on Global Commerce

A three-second autoplay counted the same as a prime-time World Series commercial, and that measurement choice broke the economics of attention, argue Joe Marchese and Jonathan Bing in an excerpt from their book Human Dot AI: How the Fight For Your Attention Became the World's Biggest Business.

The pivot point came in April 2016. BuzzFeed live-streamed two staffers wrapping rubber bands around a watermelon until it burst; the stream drew more than 3 million views, and CNBC called it the future of television. The authors, both veterans of the industry, say the label missed the core defect: nobody could tell how much of that audience actually watched, and advertisers still counted every three-second view as an impression — the industry's basic unit of value.

At the time, TV advertising was already under strain. Millions of households were canceling cable, ratings were slipping, and advertisers were moving billions of dollars to digital platforms that promised larger audiences at lower cost.

How did a subprime market for attention form?

Attention is finite — 24 hours in a day, and only so many things a person can look at at once. The algorithms that run digital advertising were not built to reflect that limit. They received a single instruction: maximize ad revenue. They learned to produce whatever generated the most impressions, counted each time an ad or surrounding content appeared on a screen for a few seconds, whether or not anyone noticed.

When genuine attention ran short, the systems expanded into inventory that required no attention at all: intercepting existing demand and claiming credit for purchases people would have made anyway. Supply exploded. Prices fell for every seller. The result, the authors write, was a flood of cheap, low-quality impressions — "a subprime market for human attention."

Television, forced to compete on those terms, ran more ad minutes per hour, which degraded the viewing experience and likely pushed more viewers toward streaming and ad-free platforms. The authors call it a vicious cycle: the less attention was worth, the more impressions the industry had to manufacture to close the gap.

What does this change for networks and buyers?

Some networks tried to break the cycle by cutting ad loads and redesigning the commercial break. Jonathan Bing, then in charge of advertising at Fox Networks Group, tested six-second spots, premium slots at the start and end of breaks, and single-sponsor streams with minimal interruptions. Other networks pursued their own versions, promising viewers fewer, better ads and advertisers more engaged audiences.

The formats offered a better value exchange for attention, but they were a hard sell. Buyers asked why six seconds should cost more than 30. Real estate and retail had long paid premiums for placement and scarcity; in advertising, brands wanted volume. Real change required advertisers to let their agencies take risks, and most refused.

Why AI raises the stakes

A decade on, the constraint has not moved — still 24 hours per person per day — but artificial intelligence can now generate unlimited content. Feeds are filling with AI-generated video, platforms automate every step from writing the ad to placing the buy, and ads now appear inside AI assistants. ChatGPT says it is on track to generate $1 billion in annual advertising revenue, with more than 100,000 advertisers running campaigns on the platform.

The risk, the authors argue, is that an endless supply of content chasing finite attention makes each impression cheaper and less meaningful than ever. The subprime market grows until advertising looks less like a market for attention and more like a toll on global commerce — while the things hardest for machines to measure, such as culture, storytelling, trust and human connection, are neglected.

The authors close on a conditional note of optimism: AI could instead be built to serve people, helping them manage time, filter information and reclaim attention. Whether the industry takes that path, they write, is a choice still available.